According to the People's Daily, on July 10, the world's first AI-native credit card, the "Kimi Credit Card," was officially launched. This project was jointly developed by Kimi, American Express, and China Agricultural Bank. It started in April this year, opened for reservations on June 12, and finally finalized the cooperation. As the first consumer voucher that integrates AI services into its core benefits system, the launch of this credit card marks a key step in the integration of AI large models from simple software subscriptions into the financial ecosystem, officially opening up a new path for intelligent payment and AI computing power consumption.

The core highlight of this credit card is achieving a "consumption as productivity" value loop. Daily spending points of cardholders can not only be exchanged for Kimi membership benefits at an 80% discount but also have a redemption ratio of 1000:1 in the points mall, with a maximum of 50% redeemable through points. Every transaction can be converted into Kimi computing power quotas or directly used to exchange for Agent (intelligent entity) usage quotas and advanced function permissions. In the future, the three parties will further deepen their cooperation and explore cutting-edge application scenarios such as intelligent entity automatic payments, Token quotas linked to credit limits, and bidirectional integration of points and computing power.

From a market strategy perspective, American Express covers white-collar workers, cross-border professionals, and business elites—core paying groups for productivity AI services. Kimi, leveraging American Express's international financial network, is expected to significantly reduce customer acquisition costs in overseas markets and accelerate its global user expansion. This cross-industry collaboration between "AI + finance" not only injects differentiated digital asset benefits into the traditional credit card industry but also provides a new traffic entry point and monetization model for the commercialization of large models.